objects hanging around after 1st Full GC

[Shane Cox]

I'm observing a strange pattern in my GC logs:  the first Full collection
doesn't reclaim much space, then the second Full collection reclaims a large
percentage of Tenured space.  This pattern repeats itself over and over
again.  Here is an excerpt from my GC logs:

First Full GC only reclaims 142,813K:
910.866: [GC [PSYoungGen: 123712K->5136K(124544K)]
1168575K->1051327K(1173120K), 0.0112247 secs] [Times: user=0.04 sys=0.00,
real=0.01 secs]
911.139: [GC [PSYoungGen: 123152K->3976K(124608K)]
1169343K->1051797K(1173184K), 0.0108060 secs] [Times: user=0.04 sys=0.01,
real=0.01 secs]
911.150: [Full GC [PSYoungGen: 3976K->0K(124608K)] [PSOldGen:
1047821K->905008K(1048576K)] 1051797K->905008K(1173184K) [PSPermGen:
26536K->26536K(53248K)], 2.1919067 secs] [Times: user=2.20 sys=0.00,
real=2.19 secs]
913.640: [GC [PSYoungGen: 118144K->6456K(124608K)]
1023152K->918339K(1173184K), 0.0300898 secs] [Times: user=0.11 sys=0.00,
real=0.03 secs]
913.881: [GC [PSYoungGen: 124600K->6464K(113088K)]
1036483K->926045K(1161664K), 0.0242545 secs] [Times: user=0.09 sys=0.01,
real=0.02 secs]

Second Full GC reclaims 796,079K:
933.943: [GC [PSYoungGen: 123280K->4736K(124992K)]
1167463K->1049778K(1173568K), 0.0099996 secs] [Times: user=0.04 sys=0.00,
real=0.01 secs]
934.247: [GC [PSYoungGen: 123520K->4240K(124864K)]
1168562K->1050930K(1173440K), 0.0099813 secs] [Times: user=0.04 sys=0.00,
real=0.01 secs]
934.258: [Full GC [PSYoungGen: 4240K->0K(124864K)] [PSOldGen:
1046690K->250611K(1048576K)] 1050930K->250611K(1173440K) [PSPermGen:
26537K->26537K(61440K)], 1.0611080 secs] [Times: user=1.07 sys=0.01,
real=1.06 secs]
935.516: [GC [PSYoungGen: 118784K->6073K(122816K)]
369395K->258744K(1171392K), 0.0122732 secs] [Times: user=0.05 sys=0.00,
real=0.01 secs]
935.704: [GC [PSYoungGen: 122809K->7164K(123904K)]
375480K->262140K(1172480K), 0.0124191 secs] [Times: user=0.05 sys=0.01,
real=0.01 secs]

Then the pattern repeats.


It's as if many of the objects require two Full GCs in order to be
reclaimed.  Any idea what's going on or how to further investigate the cause
of this behavior?

Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20090814/482e2609/attachment.htm>